OBD1 code 12?

I show code 12 as "System normal". Here's the list I looked at:

https://www.2carpros.com/articles/odb1-isuzu-1985-1994-cars-and-trucks-code-retrieval-procedure-and-definition

Is there a problem you're trying to solve? The "tach signal" is pulses from a sensor to indicate engine speed. It may be called a "camshaft position" signal or "crankshaft position" signal by other manufacturers. Besides engine speed, camshaft and crankshaft position sensor signals are also used for timing of injector pulses and firing the ignition coil(s). On some models, when one signal goes missing, the Engine Computer will enter a backup strategy that relies on the other signal, but with reduced performance. On other models, if either signal is missing, the engine will stall and not restart. On still other models, if one sensor fails while driving, the engine will continue to run normally until it is stopped, then it will not or may not restart.

Most commonly, on older engines with just one engine speed sensor, if it fails, you'll have a crank / no-start condition. Without that timing signal, the computer won't know when to fire an ignition coil. On many models, Chryslers in particular, the signal pulses from the crankshaft position sensor and the camshaft position sensor are needed to tell the Engine Computer the engine is rotating, (cranking or running), then, in response, it turns on the automatic shutdown, (ASD) relay. That relay sends current to the ignition coils, injectors, fuel pump or pump relay, and many other places. Other manufacturers do it in a similar manner, but the result is the same. No signal pulses; engine won't run.

These position sensors come in two designs, a two-wire or a three-wire sensor. Two wire sensors are just a coil of wire wrapped around a magnet. As a tooth or notch on a rotating part passes by the magnet, it disturbs the magnetic field. That induces a voltage pulse that's sent to the computer or ignition module. As with anything that generates a voltage with a magnet, speed is an important consideration. The engine must be rotating fast enough for the sensor to generate a signal high enough to be read by the computer.

Three-wire sensors have a ground wire, a power wire, typically 5.0 volts, but possibly 8.0 or 10.0 volts on some models. They still use the magnet and generate the voltage pulse the same way, but they have additional circuitry built in to amplify that signal. These put out a very nice, clean square wave signal that provides extremely precise timing.

Two-wire sensors were more common on '95 and older models and were often found in the distributor. The signals were closer to a sine wave with much less precision. Those signals were consistent though, and spark timing was adjusted with the distributor. Three-wire sensors are much more precise making them the better design for the tighter emissions rules for '96 and newer models. Spark timing is adjusted by the computer. Since it can't actually advance the spark timing to occur before the signal shows up, they start with a very high spark advance, then the computer calculates the amount of delay needed to achieve the correct spark timing. For example, base timing might be around 45 degrees before top dead center. Based on engine speed, the computer could wait for 40 degrees, then fire an ignition coil. That would result in spark occurring 5 degrees before top dead center.

Forget the idle air control valve, (IAC), (automatic idle speed motor, (AIS)), or whatever name is used by other manufacturers. That unit simply controls the volume of air allowed to bypass the throttle blade. At the same time the Engine Computer adjusts the length of time the injectors are pulsed open to control the amount of fuel. The only purpose is to adjust idle speed, not idle or running quality.

The EGR valve doesn't open at idle or low engine speeds or when the engine is cold. It should only open when its actions won't be noticed, typically at highway speeds. It is possible for it to cause rough idle when a chip of carbon breaks off and prevents the valve from closing completely. A common way to identify that is to loosen or remove the valve assembly, then reinstall it with a thin piece of sheet metal or shim stock to block the port. If the rough idle can still be made to occur, it's not due to that valve.

Oxygen sensors also have nothing to do with rough idle. Based on its readings, the computer adjusts the amount of fuel being sprayed in to keep the air / fuel mixture as close to perfect as possible. Those readings can only result in a change of about plus or minus ten percent from the default, or normal values. Incorrect operation results in increased emissions and possibly a little black smoke from the tail pipe.

The only common, but still rare problem a throttle body can cause as far as engine performance is a vacuum leak around the throttle blade shaft. That could result in erratic idle speed, just like any other vacuum leak would cause. Other than that, there are no moving parts or anything to wear like there were in carburetors. Some engines were prone to surging idle speed when carbon deposits built up on the back side of the throttle blade, but we don't see much of that anymore, in part due to better fuel additives.

A couple of thoughts come to mind for the symptom you described. While this doesn't match exactly, consider checking the ignition timing with a timing light when the engine is at a steady, low idle speed. The exact value isn't the issue. Watch the timing mark to see if it is bouncing around, especially a good ten degrees or more. If you see that, check for worn bushings in the distributor that allow the shaft to wobble around. When the magnetic pickup coil is in there, the timing pulses will vary and bounce around. For most of those pickup designs, the air gap is rather critical. A wobbling shaft can let the toothed reluctor wheel move away from the sensor's magnet. A timing pulse can fail to be big enough to be read by the computer. Lacking that pulse, the ignition coil won't fire for that cylinder. A trick to help identify that is to trigger the timing light from all of the spark plug wires, one at a time. If you only use cylinder number one, that one might never drop out, but say you use cylinder number three's spark plug wire, that one might show intermittent loss of ignition pulses when the rough running occurs.

While that last problem can cause intermittent loss of spark to individual cylinders periodically, you can also have loose or sloppy timing chains on many engines. That also allows erratic idle speed due to spark timing changes, by not so much a rough idle. Your engine uses a timing belt, so that is not likely to cause this problem.

If this were to occur right after very high-speed operation, such as with racing engines, suspect hydraulic lifters that have pumped up due to "valve float". That can result in a couple of valves being held open until those lifters bleed down. That can take as much as 15 to 20 seconds. That lifter pump up doesn't occur under normal engine speeds unless a valve spring is broken or very weak.

Probably the better place to start looking is at fuel pressure. For most engines with throttle body fuel injection, a fuel pressure regulator is built in that adjusts pressure through a vacuum hose connected to intake manifold vacuum. Two forces act on a molecule of fuel when it leaves the tip of the injector, that's fuel pressure and intake manifold vacuum. During periods of high load or acceleration, vacuum goes down, so fuel pressure is increased to offset that. Where problems occur more often is during coasting or when engine speed or load goes down. During coasting, vacuum is high. On its own, that would tug the fuel out of the injectors too aggressively, resulting in a rich coasting condition with greatly increased emissions. To prevent that, the vacuum hose to the fuel pressure regulator relaxes its internal spring, lowering fuel pressure. The net difference between the two forces, vacuum and fuel pressure, remains steady. The problem comes from that relaxed pressure regulator. The fuel pump moves a really large volume of fuel, say a gallon per minute, to the engine. Well over 99 percent of that fuel flows through the pressure regulator, then right back into the tank through the fuel return line. When the high vacuum signal tells the pressure regulator to relax, it is much easier for the gas to get through it, so the fuel volume being moved by the pump goes way up. (Less fuel going into the engine, but a lot more going through the pump), and more fuel volume that needs to get through the pickup sock, or screen, under the fuel pump housing. When that screen becomes blocked, or collapses after a few minutes of driving, insufficient volume can get through. That results in drastically reduced fuel pressure, often to the point not enough fuel gets into a cylinder for it to fire properly. Most often this causes stalling when coasting down from highway speed, and the engine runs best during acceleration and high=speed driving, but we still look at low fuel pressure for rough running complaints. Fuel filters should be at the bottom of the list of suspects when the symptoms only occur when the engine using very little fuel.

I'm going to share one more thought although I don't think this is the issue here. Based on multiple repair attempts so far, there's a good chance the owner has disconnected the battery a few times. Once reconnected and the engine is started, all sensor personalities are relearned right away, but often idle speed needs a specific set of condition for the computer to relearn when it must be in control of idle speed. This has been an extremely common concern among Chrysler owners who usually become confused when the low idle speed problem seems to fix itself. In reality, a very specific set of conditions must be met for "minimum throttle" to be relearned, but that requires driving the vehicle. For most other car brands, we don't usually hear about these problems as they have different strategies for doing these relearns. My reason for bringing this up is this may be one of those cases where at some point, idle speed goes way too low, then it takes some time for the computer to react and recover that minimum speed. You may not be feeling roughness due to cylinders cutting out. It may just be from the engine idling to slowly. You didn't say what, "a but" is as far as how much time the rough idle occurs, but the way I'd start looking at this is to be ready at the throttle body, and as soon as the roughness occurs, raise idle speed manually and see if that roughness clears up. If it still doesn't feel right, start with the fuel pressure test. If the EGR valve uses vacuum or electrical controls, unplug everything going to that valve so it won't open. Now if the roughness doesn't occur, look for carbon blocking the sealing of the valve.

Related to EGR problems that used to be common on Ford engines was a tube plugged with carbon. Rather than sending EGR gas into the intake manifold at one point, Ford sent it down a long tub with branches that fed each cylinder. As those individual branches became plugged, those cylinders ran especially well because all they got was fresh, clean fuel and air. The problem was the system was based on volume, so as more and more tubes became plugged, more of that volume went to the cylinders with tubes that weren't plugged yet. Eventually the last one or two cylinders were getting all of the EGT gas which was so high, it prevented normal combustion. This still should not be occurring at low engine speeds. It caused misfires at higher speeds when the EGR valve was expected to open.

See if those ideas get you anywhere.

Remember, the spec for timing isn't the issue. You're looking to see if it's bouncing around or cutting out.

For ignition timing, the only other thing would be a sloppy timing chain, but as I recall, yours uses a belt. With a sloppy chain, ignition timing would have to be triggered from a camshaft position sensor to make the timing vary a lot. Regardless, two degrees is fine. We need to look somewhere else.

The last thing I can suggest requires a scanner with "record" capabilities. I have a Snap-On Solus Edge and a Chrysler DRB3 for my vehicles. Many others can do the same thing. The idea is to take the scanner on a test drive, then, when the intermittent problem occurs, you press the "Record" button. Because the data goes through the scanner's memory, the recording actually begins a couple of seconds before you pressed the button, so it includes the beginning of the event. This works the same way in the shop if that's when you can make the symptoms occur.

Later, you can play the recording back slowly, frame by frame, to see what changed. There's two things to look for. The first are changes to the inputs. That might be a dropout in the crankshaft position sensor's signal, or a sudden change in a sensor's signal voltage, especially from MAP, mass air flow, and other sensors providing data used to calculate fuel needs.

The second thing to look for that can help is the outputs. You already know one of them; it's engine speed. Another common one is oxygen sensor readings. After the computer calculates fuel needs and desired idle speed, the front oxygen sensor readings tell the computer how close it came to perfect, then the computer makes tiny adjustments.

Another one to look at that's related to the oxygen sensor is short-term fuel trim numbers. Those bounce around constantly, but you're looking for a sudden change much bigger than normal, then you have to figure out why that occurred.

Something else to keep in the back of your mind is the injectors and ignition coils are fired in different ways by different manufacturers, and even differently between engine models. With three ignition coils, each one fires two spark plugs at the same time. One fires at the top of the compression stroke, and the other fires at the top of the exhaust stroke. That one is called "the waste spark". If you have a weak coil, or anything that causes no waste spark to occur, you won't miss that one, but you will get a misfire from the other one. Also, if spark is okay, a cylinder that goes lean for a few seconds, such as from low fuel pressure, the flame front is harder to keep going through the entire burn. The fuel molecules are too far apart for the flame to propagate, or jump from one to the next. Even though spark is okay, you'll get a misfire that might recover in a few seconds. This would be a case of looking at two of the four things needed, meaning spark and fuel, and ignore the things that relate to timing of those events, meaning the crankshaft and camshaft position sensors. If the scanner has graphing capabilities, you can watch when the rough idle occurs to see if the timing signals are moving around, or if they're cutting out completely. Those are two very different things with different causes to look for, but with the same result, rough idle.

A weird one I ran into years ago involved a four-cylinder engine with two ignition coils. The timing belt and sprockets were perfect, but the dowel pin between the sprocket and camshaft had sheared off, then the camshaft had slipped over time just to the point equivalent to the belt jumping one and a half teeth. At two teeth, the Engine Computer was designed to shut the engine down to protect the valves, but it hadn't quite reached that point yet. Instead, the computer was right on the edge of intermittently not firing one of the coils. The engine usually stalled from low idle speed before the computer could adjust for it, then the engine was hard to restart. At times it seemed to run fine.

For the injectors, some are fired sequentially, meaning one fires just as the intake valve right next to it is starting to open. With that design, you can have one injector misfire by itself. On other designs, they use "batch-fire". With that all three on one side of the engine fire at once. There's fewer circuits in the computer, so it saves money. The fuel sprayed from two injectors floats around in the manifold until its valve opens. A single failure in the computer or the wires to those injectors can cause all three to cut out intermittently. You'd lose half the cylinders which is a lot more noticeable than losing just one.

Another unusual one to consider is rocking of the engine between accelerating and coasting. I found a loose connector in the harness running from the body to the engine, shortly after that engine had been replaced. When the engine rocked during normal acceleration, the connector pulled apart just enough to break the connection for one or two injectors. The engine ran fine during steady speeds. A clue there was the engine had to be under load to rock. Just raising and lowering engine speed in the shop wouldn't make it act up. This defect would not have been monitored by the Engine Computer, so it wouldn't show up on the graphing scanner, but you would see the sudden lean condition from the oxygen sensor. That engine was pre-1996 so it had just the one oxygen sensor, so it took just a few extra minutes to figure out which side was acting up.

If you don't have a scanner, check at some local auto parts stores for one you can rent or borrow. This could be expensive though. In my city, they make you buy the tool, then you get a full refund when you return it. For a scanner, that could be a couple thousand dollars. If you choose to keep a tool, you still return the borrowed one, then they give or order you a brand new one. Another alternative is to find a mechanic friend who has his own, and borrow that.

By the way, for fuel pressure, be sure to not just look at the pressure. Watch what it does as the engine is slowing down and when the rough running occurs. You may need to drive the vehicle with the fuel pressure gauge clipped under the right wiper arm so you can see what the normal range is at different engine loads. Some engines, GM truck engines in particular, are very intolerant of fuel pressure just a few pounds low. Others, like on my '88 Grand Caravan, can drop from 45 to 20 psi and still run fine with no symptoms. It finally started to sputter at 15 psi. The pressure dropped over about 20 seconds when pulling a huge trailer, and the pickup screen in the gas tank was plugged. I started to describe that in a previous reply. That engine was very forgiving of low fuel pressure. Those are two extremes, but all engines will start to misfire at some point as the pressure drops.

I'm here most evenings. Take your time, then let me know what you find.